‘You Can’t Delete a Memory’: Managing the Data Past on Social Media in Everyday Life

Abstract

This article explores how the data past on social media, in the form of packaged ‘memories’, is managed by people in everyday life. Drawing on interview and focus group data, I examine how people make sense of data as ‘memories’ and how these are negotiated and managed when considered painful, awkward, or simply ‘out of place’. As such, the article outlines and discusses three specific ‘tactics’ used to manage the data past in everyday life. First, I explore the use of ‘deletion’ and how it foregrounds ways in which people seek to render certain aspects of their data past invisible, especially memories considered painful or inconsistent with the current view of self. Second, through the tactic of ‘delaying’, the article examines how some participants sought to delay emotional engagements with digital memories rather than to delete them. Finally, the tactic of ‘linking’ highlights the ways in which people sought to make sense of data as memories that were insufficiently contextualised, disjointed, or that felt simply out of place. As such, the article contributes to our understanding of the ways in which people make sense of data as well as some of the complex dynamics inherent in contemporary digital memory work.

Introduction

The rendering of human lives and practices into data through digital technologies is a mundane and normalised tendency of contemporary life. The capacity of digital technologies, platforms, apps, and algorithmic systems to generate informational patterns and correlational knowledge about societies and individuals has resulted in what has been called ‘the data revolution’ (Kitchin, 2014b). Scholars have long examined the various challenges wrought by the emergence of Big Data, understood by Boyd and Crawford (2012: 663) as ‘the capacity to search, aggregate, and cross-reference large datasets’ (see also Kitchin, 2014a; Mayer-Schoenberger and Cukier, 2013). Yet, datafication and Big Data should not only be understood in terms of the processes by which companies and platforms collect and analyse data about people, but also in terms of ‘feeding such data back to users, enabling them to orient themselves in the world’ (Kennedy et al., 2015: 1). As such, there has been a growing amount of scholarship that examines Big Data ‘from the bottom up’ (Couldry and Powell, 2014) as well as how people variously make sense of personal data in everyday life (Lupton, 2020). In particular relevance to this article, Deborah Lupton (2020: 94) points out that there is a need to further examine and understand the affective dimensions of what she calls ‘data mementos’, that is, the capacity of personal data to ‘archive and revive memories’ and to generate everyday encounters with one’s data past.

There has been a proliferation of memory features in our current media landscape in recent years, technologies where notions of memory, algorithms, and data are inextricably connected (Keightley and Pickering, 2014). Some of these, such as Apple Memories, are incorporated into smartphones whereas others, such as Facebook Memories, are integral features of larger social media platforms. These devices, also called ‘on-this-media’ (Humphreys, 2020), are mechanisms for storing, repackaging, and resurfacing past data as ‘memories’ back to people, often accompanied by notifications, such as ‘you have a new memory from five years ago’. There has been an increase in scholarship that critically examines the social power and politics of such automated memory devices (see, for instance, Garde-Hansen et al., 2009; Humphreys, 2018, 2020) and how they shape people’s encounters and engagements with their digital memories. For many people, memory devices have become an integral part of their everyday memory practices, that is, how they encounter, revisit, and negotiate traces of their data past. Yet, not much has been written about the ways in which people make sense of and manage digital memories that are considered painful, awkward, or simply out of place. This is where the article seeks to make an intervention. Drawing on interview and focus group data, this article outlines three specific ‘tactics’ (de Certeau, 1984) of managing the data past in everyday life; these being deleting, delaying, and linking. Ultimately, the article seeks to contribute to our understanding of some of the complex dynamics inherent to data sense-making in relation to contemporary digital memory work.

In the book The Practice of Everyday Life, philosopher Michel De Certeau (1984) suggested that everyday life was a result of the intersection between sociocultural, commercial, and political forces. De Certeau argued that while various social, cultural, and political institutions engender spaces of participation – what he called ‘strategies’ – there is still room for agency and ways of negotiating and subverting these spaces. This he called ‘tactics’. As de Certeau (1984) put it, ‘strategies are able to produce, tabulate, and impose these spaces, when those operations take place, whereas tactics can only use, manipulate, and divert these spaces’ (p. 30). In other words, while spaces of participation may have been produced by larger structures, people may still employ tactics in shaping the ways in which everyday life is experienced. Similarly, even though social media platforms have become increasingly active in everyday memory work, presenting the past in ‘digestible but detailed forms’ (Beer, 2019: 39), there is nonetheless space for people to tactically engage with and actively negotiate these digital memory spaces. This becomes crucial especially in the context of memories that are resurfaced to users within platforms that are considered painful or awkward.

The first section of the article outlines some of the empirical and conceptual nodes of the critical data studies literature, particularly with regards to personal data mining practices, as well as the digital memory studies literature. This is followed by a methodological discussion, outlining the different qualitative methods that generated the data analysed in this article. The rest of the article is then divided into an analysis of the three tactics of deleting, delaying, and linking in more depth. Deletion foregrounds how people seek to render certain aspects of the data invisible, especially data memories considered painful, awkward, or as being somehow inconsistent with the current representation of self. Through the tactic of delaying, I explore how some of my participants found deletion problematic and instead sought to delay emotional engagements with digital memories rather than to delete them. Finally, the tactic of linking highlights the ways in which people sought to make sense of data as memories that were insufficiently contextualised, disjointed, or that felt simply out of place. The analysis highlights some of the ways in which encountering and engaging with digital memories constitutes, in part, tactics of managing ‘undesirable’ memories being resurfaced by memory features.

Making sense of data

Everyday datafication and data mining practices have often been explored in relation to the various uses of self-tracking devices in contemporary culture. Datafication can be understood as the transformation of the world into quantifiable and measurable data. Although it is mainly understood in terms of the collection and analysis of data about users, it can also be understood as the processes of ‘feeding such data back to users, enabling them to orient themselves in the world’ (Kennedy et al., 2015: 1). In particular, it has been explored in relation to the practices and assumptions associated with the so-called Quantified Self Movement (Crawford et al., 2015; Lupton, 2016; Schull, 2016). Although the Quantified Self started out as a movement developed by two Wired Magazine editors Gary Wolf and Kevin Kelly in 2007, the term has since come to signify more than the actual movement itself. As Deborah Lupton (2013) points out, the term is now frequently used to refer to ‘the practices of self-tracking or lifelogging, terms used to denote collecting and recording data on one’s everyday life practices’ (p. 26). Here, the body is understood as form of ‘data repository’ (Lupton, 2018), rendered measurable and knowable and transformed into metrics through the (voluntary) use self-tracking devices, such as fit bits, smartwatches as well as clip-on devices. This in turn affords users to extract information about everything from heart rates and body temperature to posture and sexual performance, which can be used to tweak or improve certain habits. Lupton (2013) argues that the quantification of the body signals a conceptualisation of the body as a machine or algorithm, composed of inputs and outputs, whose performance parameters can be measured, tweaked, backpropagated, and ultimately optimised. This vision of the body as defined through the tracking of data carries echoes of Donna Haraway’s (1991) vision of the human as cyborg, the increasing hybridisation of human and machine. Yet, the assumptions underlying self-tracking practices also raise challenging questions about whether this notion of the body as data participates in what N. Katherine Hayles (1999) calls the ‘erasure of embodiment’, whereby complex embodied realities are flattened, reduced to mere informational patterns that can be endlessly modulated.

As Crawford et al. (2015) point out, the practices of self-tracking and the mining of data about bodies and selves are also intimately connected to aspirations and dreams of greater self-knowledge. In other words, knowledge through data enables the optimisation of one’s life through an increased connectivity between body, data, and quantification. The use of biometric data is seen here to enable objectivised insights into the body that were previously impossible, insights that supposedly reflect ‘the real state’ of the body (p. 480). This necessarily has an impact on people’s practices and self-perceptions, especially in relation to issues of health. That is, the optimisation or betterment of the human condition is seen as predicated on the tracking and interpretation of data, and to improve one must improve the data. Yet, the authors argue that wearables, while promising greater self-knowledge and agency, actually constitute a form of bodily control. As the body is measured, norms emerge as well as notions of what counts as normal and abnormal behaviour. Echoing Michel Foucault’s (2008) notion of biopower, self-tracking devices and the generation of small data can then be seen to generate avenues where the body becomes the locus of increased control and management through data. Moreover, the use of tracking devices to quantify and measure the self emerges as a form of neoliberal governmentality whereby bodies and selves are regulated and managed through constant comparison and competition (Beer, 2016). Attaining better numbers becomes a behavioural norm, whether that is in relation to sleep quality or sexual performance (see Lupton, 2015).

Yet, it is crucial that data are not conceptualised solely in de-humanised and de-materialised ways. Deborah Lupton (2018) evokes the notion of ‘data sense’ or ‘data sense making’ to foreground the various ways in which people engage with and make sense of the data that are collected about them. Rather than focussing on sense-making in terms of cognition or technical interpretations of data, Lupton argues that people’s sense-making practices, ‘how people live with their data’, are often ignored in contemporary scholarship: ‘the concept of data sense, as I seek to develop it here, brings the body back in, acknowledging that we learn in and through our bodies’ (p. 3). Here, Lupton (2020) draws on the feminist material perspectives of Donna Haraway and Karen Barad to view humans as assemblages of human and nonhuman actors, an idea she develops at length in the book Data Selves. Given the inextricable entanglements between humans and digital devices, Lupton (2018) understands the notion of ‘data sense’ as ‘the constitution of human and nonhuman sense-making’ (p. 3). She therefore focusses on human engagements and interactions with data in everyday life and how ‘making sense of personal data requires developing practices than can manage and interpret lively data to make them useful and knowable’ (p. 5).

This notion of data sense-making can also be seen to play a crucial role in terms of how to make sense of data visualisations (Simpson, 2020) as well as how numbers are ‘felt’ in everyday life (Kennedy and Hill, 2018). In Kennedy and Hill’s (2018) research, for instance, the response of participants when encountering data visualisations was not limited to merely the rational or cognitive. Instead, ‘emotions were evoked, and they played a significant role in shaping whether participants felt like looking, how they engaged and the information that they took from the visualisations’ (p. 840). This is significant as representations of data, whether that is through visualisations or what Lupton calls ‘data mementos’, they act as images that have ‘the potential to evoke empathy, pity, sorrow, shame and other emotions’ (p. 843). Similarly, this article conceptualises engagements with social media memories in terms of the emotions they may evoke – in this case, I focus on feelings of pain, awkwardness, and confusion, and how these are managed by people. In the following section, I will outline in more detail how social media memories can be understood in terms of data and how engaging with the data past encompasses various ‘tactics’ of management.

Memory features and the everyday data past

There is already considerable scholarship examining the ways everyday memory practices are facilitated and shaped by digital media, apps, mobile devices, and social media platforms (Van Dijck, 2007; Garde-Hansen et al., 2009; Hand, 2017; Hoskins, 2018; Özkul and Humphreys, 2015). Social media, apps, and mobile devices have developed into sites that do not simply afford communication and networked sociality, but also facilitate and shape people’s everyday memory practices and how people engage with the data past (Keightley and Pickering, 2014). In broad terms, the focus of the digital memory studies literature has focussed on three interlinking and overlapping areas: human agency, structural affordances and impact of platforms and technologies, and the digital memory objects themselves. In terms of human agency in memory-making, Jose Van Dijck (2007) proposed the notion of ‘mediated memories’, referring both to ‘the activities and the objects we produce and appropriate by means of media technologies, for creating and re-creating a sense of past, present, and future of ourselves in relation to others’ (p. 21). The focus on human agency and reflexivity in memory-making is also echoed through Lee Humphreys (2018) notion of ‘remembrancing’, signifying the ways in which people ‘purposefully and strategically create media traces to help them remember events and experiences in their lives within particular narratives of the self, the social context, and the broader cultural environments’ (pp. 73–74). The ways in which people actively and reflexively engage with technologies of memory in everyday life is important to acknowledge.

Yet, it is important that digital or mediated memories are not seen merely as add-ons to agency or the embodied capacity to remember. As Van Dijck (2009) argues, memories in the context of social media and digital technologies can be conceptualised as ‘amalgamations’, which highlights the inextricable entanglements of technology, culture, sociality, and embodiment in relation to memory practices. As such, scholars have increasingly explored the potential impact of the structural affordances of social media platforms and digital memory features on how people engage with the past. More specifically, while some work focusses on how social media platforms mine, order, and personalise digital memories (Prey and Smit, 2019), others examine how platforms afford the sharing of memories (Serafinelli, 2020) as well as how digital memory has become increasingly connective and networked (Van Dijck, 2010). Social media platforms, such as Facebook, have been examined in terms of how they circumscribe and repackage people’s past data into ‘memories’ that are frequently resurfaced in the present (Jacobsen and Beer, 2021). Popular memory or throwback features, such as Facebook Memories, Timehop, Apple Memories, and Snapchat Flashback, are mechanisms for collating and resurfacing people’s past data, such as social media posts, images, and videos back to them as ‘memories’ in the present. This necessarily has an impact on how the data past is circumscribed, encountered, and experienced.

In the context of Facebook Memories in particular, the production of memories is underpinned by a dual process of classification and ranking. That is, data, such as images and social media posts, are taxonomised into various ‘kinds’ of memories, which in turn informs the predictive capacity of the ranking algorithm in determining what memories users are most likely wanting to see resurfaced on their Newsfeeds (Jacobsen and Beer, 2021). Here, social media memories are not only conceptualised as past data points to be repackaged in the present, but they are also implicated in the logic of algorithms (see also Pereira, 2019). In short, these studies highlight the complexities of contemporary digital memory work and how emerging technologies, platforms, and apps are facilitating and shaping how people engage with their data past.

Third, scholars have also explored the various ways in which people engage with the data past through everyday digital objects. Notions, such as Van Dijck’s (2007) ‘mediated memories’ as well as Lupton’s (2020) ‘data mementos’, highlight the active memory work of people in everyday life. Yet, the importance of digital objects for memory work is particularly salient in the introduction to the early edited collection Save as . . . Digital Memories (Garde-Hansen, 2009). Here, digital memories can refer to:

Online mementos, photographs taken with digital cameras or camera phones, memorial web pages, digital shrines, text messages, digital archives (institutional and personal), online museums, online condolence message boards, virtual candles, souvenirs and memorabilia traded on eBay, social networking an alumni websites of archival material, blogs, digital storytelling, passwords, computer games based on past wars, fan sites and digital scrapbooks. (p. 4)

Although the list is not meant to be exhaustive, it does capture the multiplicity of digital objects that may operate as vehicles for people’s remembering and engagement with their data past. The crucial point here is that these notions invariably emphasise the affective potentiality of past data, their capacity to function as changeable conduits for meanings, stories, and memories from the past. As such, past data can be understood as constituting ‘evocative objects’ (Turkle, 2007), triggering affective states and connections. Yet, it is also important to note that data as memories do not provide a clear view of the past, but that they are rather objects to be interpreted, made sense of, negotiated, and managed.

Methodology

To explore how people to manage the data past in everyday life, the article draws on a combination of in-depth interviews and focus groups. More specifically, it draws on 26 remote qualitative interviews conducted from January to March 2019 as well as four focus groups conducted from May to October 2019. The data were collected as part of a larger project that explored the potential impact of algorithmic systems on the ways in which people remember and engage with the past in everyday life. Different qualitative methods were used to capture the diverse variety of practices, experiences, and encounters. The qualitative interviews were conducted with people using the memory feature called Timehop. With 21 million daily users (Lomas, 2018), the app is a popular digital tool for people to engage with and be encountered by their data past in the form of past social media images, posts, and shares, all of which are repackaged as ‘memories’. The app was selected because it was assumed that its user base would comprise active and intentional users. The majority of the Timehop users who were interviewed were not only active users, using the app on an everyday basis, but they also discussed their use of various other popular memory features, such as Facebook Memories, Apple Memories, and Snapchat Flashback, which provided points of comparative analysis.

The focus groups, on the other hand, were conducted with people with varying degrees of familiarity with social media and memory features, ranging from those unfamiliar with these features to those regularly using them. In the focus groups, participants discussed their experiences of various features, such as Facebook Memories and Apple Memories. These two features in particular were chosen because of their ubiquity, being widely embedded in people’s everyday lives. Discussing the use of the features Facebook Memories and Apple Memories provided a better understanding of the more passive ways of engaging with social media platforms and memory features as well as the experiences of being reminded or encountered by digital memories through these features (Özkul & Humphreys, 2015). It was assumed that since there is a heterogeneity of memory features currently in circulation it would necessarily engender a variety of experiences, affects, and practices. As such, one of the reasons for the focus groups was to juxtapose various memory features and their different technological affordances to provide a more nuanced understanding of the different ways in which the data past comes to matter in everyday life. Using different qualitative methods provided a more comprehensive understanding of the everyday practices and experiences of using memory features or being encountered by one’s data past through these features.

Participants for the qualitative interviews were sampled and recruited from Twitter. Twitter was used because Timehop memories are regularly shared on this platform along with a hashtag. From January to March 2019, I therefore made regular searches on Twitter for tweets containing words, such as ‘Timehop’ and ‘Timehop memories’. Potential participants were subsequently contacted directly on Twitter and invited to do an online interview to discuss their use of the memory feature and any other ones they might also use. The aim was to get a broad and diverse sample to capture as much nuance as possible when it came to remembering with social media and digital devices. Twenty-six people agreed and were provided with an information sheet and a consent form via email. The participants who agreed to take part in the research project were then provided with an information sheet and consent form to fill out before the interview. The sample was varied, both in terms of demographics (England, Ireland, the United States, Canada, and Costa Rica) and in terms of age (the sample ranged from 22 to 60). The remote interviews, conducted over Skype, lasted on average between 30 minutes and 1 hour. The focus groups, on the other hand, involved a broader sample frame and were sampled between May and October 2019. Using purposive sampling, the focus groups were recruited through advertising, emailing lists, and directly approaching local universities as well as social and community groups. The sample was considered sufficiently varied in terms of age, ranging from 18 to 70. In terms of how they were conducted, the interviewer spent a few minutes at the start explaining some of the key functionalities of features, such as Facebook Memories and Apple Memories, providing visual aids, such as images and screenshots of the features’ interfaces for clarity and intelligibility. The discussion that ensued was mostly a dialogue between participants about what they thought about the features, with some also drawing on their own experiences of using them and encountering social media memories on platforms, such as Facebook. This provided interesting analytical points of contrast.

Both the qualitative interviews and the focus groups were audio recorded and transcribed, and the participants have been pseudonymised. The interview and focus-group data were coded manually. I used thematic analysis as a way to identify patterns and themes in the interview data with regards to how people remembering with social media and digital devices. Throughout the analysis stage, categories, such as ‘practices’, ‘affects’, ‘memories’, ‘role of feature’, and ‘perceptions of the feature’ emerged. Through different qualitative methods, I was able to analyse a variety of ways in which people encounter, engage with, negotiate, and manage their everyday data pasts through contemporary memory features. The following sections outline three principal ways or ‘tactics’ through which people could be seen to manage their data past, deletion, delaying, and linking. These tactics are not meant to be exhaustive, but rather provide insights into some of the complexities and nuances of what constitutes memory work in the age of social media platforms and datafication.

‘A little pre-emptive damage control’: deletion

One of the effects of the proliferation of social media platforms and memory features is that it is increasingly common for people to encounter painful or awkward digital memories. That is, certain posts, images, or videos referring to painful events or relationships, past data resurfaced at an awkward time in a person’s life. One tactic used to manage these undesirable memories was through acts of deletion. For one of the participants, memory features, such as Timehop, were useful for what he calls ‘a little pre-emptive damage control’ (Francis). Similar to Facebook Memories, Timehop affords users the option to automatically cull data from multiple social media platforms and devices. As a result, when using the app people encounter diverse ‘memories’, such as old Facebook posts, tweets, images from Instagram and Snapchat, or videos stored on one’s smartphone (Timehop, 2019). As Timehop resurfaced past data as memories, it allowed Francis to go back and ‘delete some of the more cringy things I might have put on social media or put out there’. Yet, this was mostly conceived as ‘damage control in an innocuous sense’ as it chiefly related to corny jokes from high school. Another participant would similarly only delete memories ‘for weeding out meaningless content that I don’t really need to see every year’ (Jack). In these cases, the use of deletion was understood more as an act of curating, organising, or simply tidying ones digital photo galleries. A form of mnemonic weeding out’.

However, there were also cases where people would encounter aspects of their data past that were less innocuous, so to speak. Miriam stated that she would delete memories if they no longer seemed ‘appropriate’, including photos and posts she did not ‘agree with anymore’. In these cases, deletion constituted a way to manage and tweak past presentations of self, ensuring that such memories would not be resurfaced in the future. This was also echoed in another interview where the participant discussed how he, some years ago, had taken a ‘very short-sighted point of view’ when posting about some personal opinions (Oliver). As this past post now resurfaced as a memory on Facebook, it represented a disjoint between his past and present perceptions. As such, Oliver stated that ‘I’m going to eliminate this’, since ‘I no longer believe that’s true anymore’. In these cases, deletion constituted not only as a technology of self in terms of tweaking self-representations, but also as a tool for reflection. A mechanism to ensure narrative consistency and to avoid further painful encounters with certain personal data. Deletion can therefore be understood as a tactic of managing the data past, negotiating and making sense of past data points in light of the present, eradicating painful or awkward memories that no longer fit the current trajectory of the self. As a tactic of everyday data management, deletion constitutes highly reflexive ways of revisiting, assessing, negotiating, and making sense of one’s data past.

In other cases, the emotional impact of seeing memories automatically resurfaced on social media and digital memory devices can be too much. One of the participants explained how one of her close friends had deleted Timehop from his phone:

A friend of mine he told me that he stopped using it at some point, because it would bring up too many negative memories, he had old pictures. His girlfriend had passed away, and it was just too hard to have these memories pop up, you know the pictures that you’ve saved on his phone that he wants to keep or just the regular Facebook posts that he had posted in 2009 or whenever, happy times with him and his girlfriend and to have that come up it was just too hard. So he didn’t want the negative memories, so that’s why he stopped using Timehop. (Imogen)

Whereas for some of the participants, it was enough to delete certain memories, to render them invisible and unresurfaceable, Imogen explains how seeing some memories brought forward by digital media can result in a complete opting out. As a result, she adds,

It makes it very easy to scroll past stuff, but I can imagine that if it’s pulling up stuff from a painful relationship or my friend his loved one, who had passed away, some of those things might be if you’re not quite ready for it can be tough to see and deal with.

Yet, it is important to note that deletion in this context is not understood as the complete eradication of data; rather, it is a means of rendering certain aspects of one’s data past invisible, ensuring that they do not resurface in the future. When users ‘delete’ their data off of social media platforms or online databases, it does not necessarily mean that these are fully deleted. In fact, it is questionable whether deleting one’s data from social media platform is ever fully possible (Hetherington and Lee, 2000), especially since people’s digital traces continue to be economically valuable to larger companies and infrastructures. As Nanna B. Thylstrup (2019) points out, ‘what makes digital traces valuable, then, are the infrastructures that subtend the processing of data points, as well as the retemporalisation of these traces from past to future, making data actionable in new ways’ (p. 3). Similarly, social media platforms have long been vast repositories or archives for people’s past data (Garde-Hansen, 2009). They increasingly encourage engagement with the data past: previous times, moments, and events as stored on and engendered through the platform. The tactic of deletion figures here as a way to negotiate social media platforms as archival structures, as repositories for mediated memories. At the same time, however, social media platforms problematise the efficacy of deletion as a means of negotiating the data past in everyday life. In short, deletion highlights the tensions between the everyday tactics of people and the larger digital structures and spaces that circumscribe memory work.

‘Maybe next year I’ll reread this memory’: delaying

Delaying figured as another salient tactic for managing undesirable memories being resurfaced. When asked how she reacts to seeing painful or awkward memories crop up, one participant stated that ‘I just try to scroll past it as fast as possible, because I’m like, oh I don’t want to be reminded of that’ (Miriam). This was also echoed by another participant who stated that ‘I’ll give myself permission to tap very quickly through it and not engage with it, and just get to the end and close the app and be done with it’ (Diana). When asked if she still has any emotional connections to those digital memories, Diana responded,

Diana:

Only choosing not to have an emotional connection that day. Not today, maybe next year I’ll reread this memory

Interviewer:

Is there a particular reason for you maybe not to have an emotional connection to some of the photos?

Diana:

Yeah, I think it’s more okay I know what just happened, I will deal with this emotionally at the time and place of my choosing, and it’s not today.

As this suggests, encountering and engaging with digital memories that are resurfaced on platforms, such as Facebook, or features, such as Apple Memories, involves conscious decisions when to emotionally engage with certain memories, especially if these are somehow painful, uncomfortable, or awkward. By delaying the memory, so to speak, engaging with it at a later point in time, the participant can be more emotional prepared to ‘deal with this’. Although this tactic raises questions to what extent people are able to always detach themselves emotionally from such mnemonic encounters, it does highlight delaying as a way in which people may manage and circumvent certain uncomfortable aspects of their data pasts.

This tactic of delaying an emotional engagement with certain digital memories resonates with calls for ‘slowing down’ in a culture dominated by the logic of speed (Rosa, 2013). In their book Slow Computing, Rob Kitchin and Alistair Fraser (2020) argue there is a need for a new paradigm shift in contemporary culture in terms of how people understand and relate to computing and digital devices in everyday life. Their argument is two-pronged, emphasising both how people interact with everyday devices and a call to ‘step back, even if just a little, to try and seize some self-control’ (p. 2). They propose the notion of ‘slow computing’ to, as they state,

Redirect attention, reclaiming time for other pursuits, and to protect its practitioners from any pernicious effects of living a digital life. It does this both personally, by providing tactics for managing how you engage with a digitally mediated world, and collectively by pooling knowledge, cooperatively acting together, and mobilizing political power to shift public debate and influence the regulation of questionable practices. (p. 3)

The tactic of delaying can be seen in similar terms. It can be understood as a form of ‘slow computing’ which provides an everyday tactic for managing how people encounter and engage with digital memories being resurfaced on social media platforms and memory features. Similar to Kitchin and Alistair’s notion, delaying constitutes a way to resist, or least to problematise, the regime of speed and constant acceleration that underpins the current digital age (Gane, 2006). In short, delaying can be understood as a means to try and seize some control over one’s own data past.

The tactic of delaying, however, was seen by some participants not only as a means to manage certain aspects of one’s data past. Rather, it also figured as a way to circumvent the need for deletion. One of the reasons for delaying emotional engagements with certain digital memories seemed to be that many of the participants I interviewed found the idea of deleting memories disconcerting and problematic for various reasons. One reason was well encapsulated by one of the participants, who stated that,

The only time I ever delete photos is if it’s a blurry photo or a screenshot I don’t need anymore. I don’t like to delete anything . . . every story has some sort of focus or point or story or something that when you delete a photo, whether or not it’s five photos in a row of the same thing, each one can evoke a different memory, can evoke a different feeling. I don’t want to lose that. That’s the whole point of taking photos is that you want to remember everything. (Keith)

In this case, deletion is seen as tantamount to a loss of memories, a loss of one’s own self in a way. Delaying, in contrast, figures as a way to avoid the decisions necessitated by deleting certain memories. It is also evident from this that deletion occupies a contested space in a context where any clear ontological limits between digital memories, data, and the self are problematised. This is alluded to in another interview where they stated that,

My photo gallery on my phone has actually been left intact for Timehop, and it has caused me a lot of problems because I have past relationship photos in there, and my girlfriend will say to me why don’t you want to delete these? I have built the concept that if I eradicate these, I’m sort of eradicating the memories. (Charlie)

Here, deletion is not just a question of simply honing or curating a certain presentation of self by discarding certain past data points, such as painful social media images or awkward tweets. For these participants, it is instead tantamount to the deletion of actual memories. As such, tactics of managing one’s data past can be seen as intimately entangled with complex issues of identity and memory.

These tensions inherent in tactics of managing the everyday data past echo one of Colin Koopman’s (2019) arguments in his book How We Became Our Data. He suggests that that the idea of deletion will occupy an increasingly contested and complex role as people are increasingly ‘swaddled in data’ (p. viii). Contemporary life is inextricably constructed and construed through data. This means that data are intimately implicated in people’s notion of agency and personhood, resulting in what Deborah Lupton (2020) calls ‘data selves’. In other words, data are intimate parts of who we are. As people increasingly make and are made up by data, the deletion of their personal data may be seen as a sort of violence, a tampering with the self, an unravelling of intimate aspects of identity. As a result, delaying figures, in part, as a way to avoid the potential negative implications of deleting aspects of one’s data past. Any decisions related to deletion and delaying are highly affective, constituting an ongoing negotiation of one’s own sense of self. How to manage one’s data past, then, is seen to be entangled with questions of how to manage aspects of oneself or narratives of self that occupy a problematic space.

‘You don’t get the chain of thoughts and memories’: linking

In the context of managing the data past, not all digital memories were necessary painful, uncomfortable, or awkward by the participants. On the contrary, there was a plethora of affective responses, feelings, and perceptions that were evoked in the interviews. Specifically, participants also discussed how they felt when some memories resurfaced that seemed out of place, that is, without sufficient context and seemingly lacking in any particular meaning and significance. The tactic used by many participants in these situations I call ‘linking’, aptly evoked in one of the interviews,

Sometimes a few things will come up and I will look at it and think what was that? There was no link . . . then you don’t know what the link is. A lot of them like yesterday it must have lots of them because I was constantly going back to the links and reading the tweets. I think it was just I was having conversations with people and I was curious as to what started the conversation off. (Lydia)

The apparent absence of ‘links’ or contextual cues for certain memories suggests that social media platforms and memory features do not resurface pre-packaged, coherent narratives about people – although they increasingly attempt to do so, either through the use of machine learning algorithms (Jacobsen, 2020) or through the structural affordances of the platform, such as the Facebook Timeline (Van Dijck, 2013). Rather, the memories being resurfaced are often decontextualised and disjointed, without sufficient context, and therefore, need to be made sense of. The importance of context in relation to memory work was also echoed in one of the focus groups I conducted, where some participants stated,

Brian:

If you’re cued during conversation, that one then leads to something else; if you’re cued off a phone, you go oh look that’s interesting! I remember x y z and then it stops

Daniel:

Exactly!

Brian:

There’s not a continuation of it’s a very short and I presume that’s why you have to do it often, because you get no chain. You don’t get the chain of thoughts and memories because it just dies itself down because it’s not being reinforced by the other person/

The participant uses the metaphor of ‘a chain’ to suggest the extent to which emotional engagements with digital memories depend on their wider contexts, the temporal positions they occupy in the linearity of the chain of events. Later in the interview, one of the focus group members, Anna, found these disjointed memories to be shallower than others, as not really being memories at all in any meaningful sense. This is particularly relevant, given the fact that social media platforms and memory features only provide what they think is a memorable image and not enough contextual information relating to it (Jacobsen, 2020).

Linking, in this context, implies active and reflexive processes by which disparate data elements are woven together into meaningful wholes to manage aspects of the data past that seem disjointed or simply out of place. One way in which this was exemplified was in one of the interview, where the participant who stated that,

In the times that I do have a bunch of photos there it is occasionally fascinating to be like to oh what was I doing on this day? . . . If I see that there is a whole bulk of activity on a platform that I don’t usually have a bulk of activity on, that’s a sign that I should be digging a little bit further to remind myself what was this. (Oliver)

As this participant suggests, the insufficiency of contextual cues or the absence of links necessitates greater involvement or ‘digging’ by the people using memory features. Memories may be repackaged, automated, and resurfaced, yet they still require work. This echoes Annette Kuhn’s (1995) notion of ‘memory work’, where remembering is seen as an active ‘investigation’ into past experiences as well as into the digital objects that mediate them. Moreover, Keith stated that,

There are some days where I’ll see a status, I’ll see a photo in my library, and I’m just like what the hell was I doing that day? I have to go down a rabbit hole almost to figure it out. I’ll go on Facebook; I’ll look who I tagged and I’m just like trying to figure out why are we there? Eventually I’ll get to it. (Keith)

As this participant suggests, the process of making sense of insufficiently contextualised data points can be akin to going down ‘a rabbit hole’. Linking here becomes tantamount to processes by which people revisit a memory’s place of origin, that is, the platform where the mediated memory originated: checking the tags, who tagged, when it was tagged, and so on. Thus, linking, or what one participant called ‘social media palaeontology’, becomes a way to manage data and making sense of memories that seem out of place, memories that are insufficiently contextualised.

The tactic of linking can also be better understood through the conceptual lens of ‘emplotment’. Derived from narrative theory, emplotment signifies the minimally necessarily elements needed to constitute a narrative – these elements being characters, action, and plot (Ricoeur, 1983). Steph Lawler (2014) deploys emplotment as a way to investigate how meanings, memories, and past experiences feature in the formation and maintenance of narratives of self. For Lawler, emplotment signifies the ‘processes of producing an identity through assembling various memories, experiences, episodes, etc., within narratives’ (p. 24) Emplotment, therefore, constitutes the processes by which memories, experiences, and past events are ‘linked together’ into a meaningful whole, a narrative (p. 24). Linking, we discussed in this article, shares many of these qualities with the notion of emplotment. It can be similarly understood as processes by which people reflexively synthesise events, episodes, data, tweets, social media images, and memories, into meaningful and wholistic narratives about self. Linking foregrounds how encountering and engaging with one’s data past is a salient form of data sense-making, that is, processes of managing and making sense of memories resurfaced by social media platforms and memory devices. Linking, more specifically, showcases how aspects of the data past, especially those memories that are insufficiently contextualised, come to matter in the present.

Yet, the tactics of linking does not amount to an unearthing of a ‘pure past’, so to speak. As was stated in one interview, ‘the story is a lot more complicated than that, but that requires a lot more digging and thought and memory, than the thing that I’m shown regularly’ (Diana). Identities do not come pre-packaged, but are rather continuously made and remade, woven with the mingled yarn of stories, meanings, and memories of the past. As a result, making sense of data (in this case, insufficiently contextualised memory objects) is akin to ongoing processes of digging, a form of labour. Through this, it becomes clear that tactics of managing the data past in everyday life are simultaneously ongoing processes and practices of memory work.

Conclusion

As social media platforms and memory features proliferate, and as people increasingly ‘become their data’ (Lupton, 2020), the management and negotiation of the data past will occupy an increasingly complex role in people’s everyday lives. For many people, memory devices and features, such as Timehop and Facebook Memories, have become an integral part of their everyday memory practices, that is, how they encounter, revisit, and negotiate their data past. Drawing on interview and focus group data, this article has explored how people make sense of and manage aspects of the data past resurfaced on these platforms that are considered painful, awkward, or insufficiently contextualised.

More specifically, I have outlined three ‘tactics’ (de Certeau, 1984) by which people practically manage their data past in everyday life. The tactic of deletion highlights the desire that people sometime have to render undesirable social media ‘memories’ invisible, ensuring that these do not resurface again at some time in the future. Yet given the ‘intimate entanglements’ (Latimer and López Gómez, 2019) between people’s memory practices, data, and notions of self, deletion was also considered by some of the interview participants to be highly problematic. The self is in a state of perpetual becoming, a process (Hockey et al., 2013). The self is also perpetually changing and evolving together with technologies and media, and as such, it is reasonable to argue that people’s relationship to the past is in a constant state of flux. The role of deletion will therefore carry multiple, contested meanings and may also change over time. One indication of the contested role of deletion is that people also deployed ‘delaying’ as a tactical response to encountering undesirable, painful, or awkward ‘memories’ resurfaced on memory features. That is, delaying constitutes conscious and active decisions to delay emotional engagements with certain aspects of the data past, or as one participant put it, ‘maybe next year I’ll reread this memory’. On one level, delaying echoes the call for a cultural paradigm shift towards ‘slow computing’ (Kitchin and Fraser, 2020), and on another level, it also figured as a way to circumvent the need for deletion. Finally, I suggest that one tactic which is also crucial for managing the data past in everyday life is ‘linking’. Akin to the notion of ‘emplotment’ (Ricoeur, 1983), linking emphasises the ways in which engaging with one’s data past is a salient form of data sense-making, that is, processes of managing and making sense of memories resurfaced by social media platforms and memory devices. This tactic was especially relevant in situations where participants encountered digital memories that were without sufficient context, felt disjointed, or that seemed simply out of place.

This article has sought to deepen our understanding of the ways in which people make sense of data as well as some of the complex dynamics inherent to contemporary digital memory work. As I have pointed out, social media and memory work have become increasingly interwoven in our current media landscape. Tactics, such as deletion, delaying, and linking, although they highlight the agency of people in everyday memory work, they also indicate the ways in which social media platforms have the power to reshape how we engage with, negotiate, and remember the past. Instead of us deciding what a memory is, we are increasingly presented with ready-made memories on social media, memories that have already been imagined, classified, ranked, and resurfaced – that is, automatically produced. In short, social media platforms circumscribe what counts as the memorable past, what should or should not be remembered. As our memories are increasingly being automatically produced by social media platforms and presented to us as ready-made objects, our relationship to our memories in social media spaces changes. Instead of memories to be mined from the archives of social media, our engagements with the data past will be a matter of tactics, of negotiating a wide variety of algorithmically resurfaced memories, some of which will be awkward, painful, and unwanted, preferably deleted or delayed.

There is therefore still more research needed into the specific ways in which social media platforms and contemporary memory features affect and shape people’s remembering of and engagements with the past. As Deborah Lupton (2018) states, ‘personal data can have agentive capacities that shape people’s embodied responses and actions, their sense of selfhood and their relationships with other people and with other things’ (p. 6). Similarly, there is a need to better understand the agentive capacities of memory features, such as Facebook Memories or Apple Memories, and how meaningful encounters with the data past are facilitated, produced, and indeed achieved through the various interplays between humans and machines in everyday life.